Research on the impact of digital finance on rural revitalization

Abstract

As an important driving force for rural revitalization, digital finance can effectively enhance the efficiency and quality of the rural economy, improve the living conditions of farmers, and promote the comprehensive progress of rural society. This paper utilizes panel data from 281 prefecture-level cities in China over the period of 2011–2022 and employs quantitative methods such as the double fixed effects model, the mediating model, and the threshold effect model to conduct an in-depth analysis and assessment of the impact of digital finance on the rural revitalization process. Research indicates that digital finance significantly promotes rural revitalization, a conclusion that remains valid after a series of robustness tests. Digital finance indirectly empowers rural revitalization by leveraging technological innovation and enhancing capital allocation efficiency. Its impact on rural revitalization exhibits a dual threshold effect across different levels of economic development. The impact of digital finance on rural revitalization exhibits regional heterogeneity, with its promotional effect being particularly pronounced in eastern regions. Accordingly, future efforts should focus on accelerating the development of digital finance, enhancing technological innovation, optimizing capital allocation efficiency in rural areas, and implementing differentiated economic development strategies to foster deeper integration between digital finance and rural revitalization.

Introduction

The Rural Revitalization Strategy is a pivotal component in building a modern socialist country in all respects, carrying the mission of advancing agricultural and rural modernization, improving farmers’ quality of life, and promoting common prosperity. Since its introduction, China’s rural development has entered a new phase. Its implementation requires prioritizing agricultural and rural development, promoting integrated urban–rural development, and vigorously advancing the revitalization of rural industries, talent, culture, ecology, and organizations. This strategy holds significant importance for strengthening the agricultural foundation and advancing Chinese-style modernization. However, implementing this strategy is a complex and long-term process facing multifaceted challenges, including resource allocation, industrial transformation, environmental protection, and social governance. How to achieve the comprehensive goal of “a strong agriculture, beautiful countryside, and prosperous farmers” has become a critical issue requiring urgent attention in both academic research and policy practice.

The realization of rural revitalization fundamentally requires both depth and breadth in funding and resources. Rural infrastructure development, agricultural industry upgrades, and improvements in farmers’ livelihoods all demand substantial financial support. Digital finance, as a new financial model leveraging big data, artificial intelligence, blockchain, and other technologies, has significantly expanded the boundaries of financial services, enhanced service efficiency and accessibility, and provided an opportunity to address the allocation of financial resources in rural areas and narrow the urban–rural development gap. Consequently, practical questions such as “Can digital finance become a driving force for rural revitalization?”, “What are its specific mechanisms of action?”, and “Does it have non-linear effects?” warrant further exploration.

From a theoretical perspective, existing research either focuses on the impact of digital finance on individual economic indicators or explores isolated pathways for advancing rural revitalization. It has yet to establish a systematic analytical framework that deeply integrates these two dimensions, particularly lacking theoretical elaboration on “how digital finance can align with the multidimensional objectives of rural revitalization.” From a practical perspective, traditional finance has long suffered from service gaps in rural areas, creating a contradiction with the urgent need for diversified financial support in rural revitalization. The technological characteristics and inclusive nature of digital finance offer a potential solution to this contradiction. Targeted research is urgently needed to clarify its operational logic and implementation pathways.

This study focuses on the Chinese context primarily due to the unique developmental characteristics of China’s rural revitalization and digital finance sectors. China’s rural revitalization has established a systematic advancement pathway integrating top-level policy coordination with local innovative practices, providing a stable institutional environment for the penetration of digital finance into rural areas. Simultaneously, within a unified institutional framework, the phased differences across regions in terms of rural revitalization foundations and digital finance penetration levels enable more precise identification of core influencing factors. This approach avoids the interference caused by differences in institutions, cultures, and economic foundations that often arise in cross-national comparisons, making the research conclusions more targeted and interpretable.

To address these issues, this paper first constructs a theoretical framework for the multidimensional impact of digital finance on rural revitalization based on financial exclusion theory, innovation diffusion theory, and gradient development theory. Relying on panel data from 281 prefecture-level cities in China over the period 2011–2022, Kernel density estimation is employed to analyze the dynamic evolution of the distribution of rural revitalization and digital finance development levels in China. Second, a double fixed-effects model is applied to accurately analyze the direct effects of digital finance on rural revitalization. Meanwhile, a mediating effect model is utilized to investigate the role of technological innovation and capital allocation efficiency in the impact of digital finance on rural revitalization and heterogeneity analysis is conducted from different geographical endowment perspectives. Finally, a threshold effect model is constructed to explore the nonlinear impact mechanisms of digital finance on rural revitalization under different levels of economic development.

The potential marginal contributions of this study are primarily reflected in three aspects. First, in terms of research content, this study constructs a theoretical analytical framework for digital finance to serve the rural revitalization strategy and conducts empirical tests based on prefecture-level panel data, which provides micro-level evidence for revealing the mechanisms through which digital finance empowers rural revitalization. Second, in terms of research dimension, it breaks through the limitations of the traditional provincial macro-analysis paradigm and innovatively shifts the research focus to the prefecture-level geographical unit, offering a new perspective and approach for rural revitalization in the digital finance era. Third, at the methodological level, this paper employs a multi-mechanism analysis and panel threshold model to offer a novel analytical perspective on the relationship between digital finance and rural revitalization. Specifically, through multi-mechanism analysis, this study identifies the intrinsic logic by which digital finance indirectly influences rural revitalization via dual pathways: technological innovation-driven development and optimized capital allocation efficiency. Compared to single-mechanism analysis, this approach provides a more comprehensive exploration of the transmission channels through which digital finance empowers rural revitalization. Building upon this foundation, the panel threshold model is further applied to explore their relationship, ultimately identifying a dual threshold effect of economic development levels. The enabling effect of digital finance on rural revitalization exhibits a nonlinear pattern of increasing marginal benefits. Consequently, this study not only deepens the understanding of digital finance’s role in empowering rural revitalization but also provides empirical support for formulating differentiated regional strategies, offering valuable insights for related research and practice.

Literature review

While existing research on the relationship between digital finance and rural revitalization has yielded substantial findings, most studies have focused narrowly on the impact of digital finance within isolated sectors—agriculture, rural areas, or farmers—rather than providing a comprehensive analysis of its integrated role within the broader rural revitalization system.

Research on pathways for digital finance to drive agricultural development

In exploring the mechanisms through which digital finance propels agricultural development, scholars widely acknowledge that digital finance has become a pivotal force in overcoming traditional financial coverage bottlenecks in rural areas and advancing agricultural modernization by innovating service models and optimizing resource allocation^1,2. First, digital finance alleviates rural credit constraints, thereby boosting agricultural productivity^3,4. Simultaneously, it enhances farmers’ willingness to adopt agricultural technologies, increasing production efficiency^5,6. Further research confirms that integrating digital technology with inclusive finance breaks financial exclusion in agricultural production. By supporting farmers’ innovation and entrepreneurship, it accelerates the diffusion of agricultural technologies, providing sustained support for agricultural development at the technological penetration level^7,8. As research deepens, some scholars have expanded their focus to agricultural green transformation and sustainable development. Findings indicate that digital finance exerts a nonlinear mitigating effect on agricultural non-point source pollution. When the non-agricultural share of farmers’ income exceeds a specific threshold, its environmental improvement effect significantly strengthens^9,10,11, providing new empirical evidence for digital finance’s contribution to sustainable agricultural development. Additionally, other studies have examined the impact of digital finance on agricultural total factor productivity. Findings indicate that digital finance can drive improvements in agricultural total factor productivity by accelerating land transfer processes and deepening agricultural technology diffusion, thereby providing crucial momentum for enhancing the quality and efficiency of the agricultural industry^12,13,14.

Research on the effects of digital finance in empowering rural development

The enabling effects of digital finance on rural development are primarily manifested in promoting urban–rural integration, optimizing industrial structures, and boosting the rural economy^15,16,17. Its operational mechanisms and impact characteristics have become increasingly clear as research deepens. In terms of specific pathways, digital finance can enhance the flow and integrated development of urban–rural factors by increasing entrepreneurial activity in rural areas^18,19. Simultaneously, it drives high-quality growth in the rural economy by improving economic efficiency and optimizing the allocation of urban–rural resources^20,21. As research perspectives have sharpened, scholars have further revealed the enabling effects of digital finance in specific rural sectors. For instance, it can stimulate rural entrepreneurship by alleviating financing constraints for agricultural operators^22,23. Furthermore, technological innovation can optimize rural energy consumption structures, thereby mitigating rural energy poverty²⁴. Notably, studies consistently find significant regional heterogeneity in digital finance’s impact on China’s rural development. Specifically, its local driving effect on rural economies is more pronounced in eastern coastal and central inland regions, while its efficacy is relatively weaker in remote western areas^25,26. This disparity provides a reference basis for formulating regionally differentiated digital finance policies supporting agriculture.

Research on the impact of digital finance on rural development

Existing studies examining the relationship between digital finance and rural development have primarily focused on the income dimension, conducting multifaceted analyses of direct effects, heterogeneous characteristics, and indirect pathways of income growth^27,28. Specifically, digital finance significantly promotes the simultaneous growth of farmers’ operating income, wage income, and transfer income. This effect stems from digital finance’s multidimensional support for agricultural production and management, as well as farmers’ employment and income growth, providing new impetus for optimizing farmers’ income structure²⁹. Simultaneously, research reveals pronounced heterogeneity in the income-boosting effects of digital finance among farming households. This effect is more pronounced among households with higher education levels, larger-scale operations, and higher mechanization rates. This indicates that human capital levels and agricultural production scale are prerequisites for digital finance to fully leverage its income-enhancing potential³⁰. Regarding the operational pathways, scholars have further revealed the indirect mechanisms through which digital finance drives farmers’ income growth: On one hand, digital finance can reduce land transfer transaction costs, promote large-scale land management, and thereby indirectly increase farmers’ land rental income and agricultural operational efficiency^31,32. On the other hand, it lowers barriers to non-agricultural entrepreneurship by providing financing access and market information resources, thereby improving the efficiency of non-agricultural operations and ultimately increasing farmers’ non-agricultural income^33,34.

Research review

Although existing research on digital finance and rural revitalization is relatively abundant, three areas warrant further exploration:

First, in terms of research perspectives, existing literature predominantly follows a “fragmented” analytical approach, lacking a coordinated examination of the multidimensional objectives of rural revitalization. Most studies tend to examine the impact of digital finance on isolated dimensions—such as farmers, agriculture, or rural areas—failing to establish a synergistic development framework that integrates multiple objectives like “prosperous industries, ecologically pleasant living environments, civilized rural customs, effective governance, and affluent livelihoods.” While this analytical paradigm reveals localized effects of digital finance, it struggles to systematically explain how resource integration and mechanism coordination drive the holistic process of rural revitalization. This paper enriches the field by constructing a comprehensive evaluation index system covering five dimensions of rural revitalization and employing a dual fixed-effects and mediation effects model.

Second, at the spatial scale, existing research predominantly relies on provincial-level macro data, overlooking the heterogeneity and policy adaptability of prefecture-level units. While provincial panel data reflects overall trends, it tends to obscure the spatial heterogeneity of urban–rural dual structures within provinces, the nonlinear relationship between digital financial infrastructure penetration and urban economic carrying capacity, and the policy iteration and adaptive adjustment effects of local governments in utilizing digital financial tools. This paper breaks from traditional analytical paradigms by lowering the research horizon to 281 prefecture-level cities. Utilizing kernel density estimation, it reveals the spatial dynamics of digital finance and rural revitalization, thereby providing evidence for differentiated policy implementation at a finer geographical scale.

Third, regarding mechanism identification, the intrinsic pathways and nonlinear characteristics through which digital finance influences rural revitalization remain under-explored. While existing research suggests digital finance may function via channels such as alleviating credit constraints and promoting technological innovation, empirical tests of its transmission mechanisms remain unsystematic, particularly lacking discussions on nonlinear relationships across different stages of economic development. This study not only introduces two mediating variables—technological innovation and capital allocation efficiency—to validate their transmission role between digital finance and rural revitalization, but also constructs a panel threshold model with economic development level as the threshold variable. It reveals the phased characteristic of “increasing marginal benefits” in the enabling effects of digital finance, thereby deepening our understanding of the dynamic impact mechanism of digital finance.

Theoretical mechanisms and research hypotheses

Direct impact of digital finance on rural revitalization

Against the backdrop of the digital wave and the rural revitalization strategy converging, digital finance is evolving from a technical tool into a core vehicle for reshaping the rural financial ecosystem.From the perspective of financial exclusion theory³⁵, traditional rural finance struggles to meet the capital demands of rural revitalization due to “geographical exclusion,” “assessment exclusion,” and “conditional exclusion.” This manifests in tangible challenges: low coverage of traditional financial outlets in townships, exclusion of numerous smallholder farmers from financial services due to lack of collateral, and protracted credit approval processes. This phenomenon aligns with findings by some scholars that “the rural financial accessibility gap constrains rural development”^36,37. Digital finance offers a systematic solution to traditional financial exclusion through technological innovation, forming the core logic of its direct empowerment to rural revitalization. First, leveraging tools like mobile banking and online payments, digital finance overcomes geographical barriers, enabling farmers in remote areas to access financial services equivalent to urban residents. This effectively addresses “geographical exclusion,” significantly improving service coverage efficiency compared to traditional physical branch expansion models. Second, digital finance constructs credit profiles using multidimensional data—such as farmers’ e-commerce transaction records and agricultural machinery usage—substituting data-based credit for traditional collateral. This alleviates “assessment exclusion” and significantly increases the proportion of unsecured loans in rural areas. Finally, the application of intelligent approval systems substantially shortens loan processing times, reducing transaction costs stemming from “condition exclusion.” As illustrated in Fig. 1, the comprehensive exclusion-alleviating effects of digital finance extend further into various dimensions of rural revitalization, specifically manifested as follows:

Fig. 1

Multi-dimensional impact mechanism of digital finance driving rural revitalization.

In the dimension of industrial prosperity, digital finance provides “short-term, small-scale, urgent, and frequent” credit support to rural business entities such as family farms and cooperatives, effectively invigorating rural entrepreneurship and forming a “entrepreneurial incentive—industrial upgrading” chain driven by digital finance³⁸. In the ecological livability dimension, digital finance uses green credit tools to specifically support eco-friendly projects like water-saving irrigation and organic agriculture, propelling agricultural production toward sustainability³⁹. In the rural cultural dimension, establishing rural credit platforms has increased loan approval rates for creditworthy farmers, with this credit incentive mechanism fostering widespread awareness of social credit in rural communities. In the governance dimension, digital identity verification and payment systems integrated into village management processes significantly enhance oversight of collective funds. In the prosperity dimension, mobile payments’ widespread adoption in rural retail scenarios drives sustained growth in e-commerce transactions, directly boosting farmers’ incomes⁴⁰.

In summary, we propose the following hypothesis:

H1: Digital finance effectively promotes rural revitalization.

Indirect impacts of digital finance on rural revitalization

The indirect empowerment of digital finance for rural revitalization unfolds through two core pathways: technological innovation-driven development and optimized capital allocation efficiency. These approaches respectively leverage technological breakthroughs and capital integration to transform digital finance’s technical advantages and resource allocation capabilities into momentum for rural development.

From the perspective of technological innovation pathways, technological innovation serves as the core driving force behind agricultural modernization. However, the transformation of technologies in rural areas faces triple constraints: funding shortages, information lags, and elevated risks. Existing literature confirms that digital finance can alleviate R&D financing constraints and promote technology transfer^41,42,43, yet most studies focus on urban industrial sectors. The long technology cycles, high risks, and strong regional adaptability of agricultural technologies make traditional finance inadequate, creating a research gap. Innovation diffusion theory provides a suitable analytical framework, emphasizing that three elements—funding, information, and risk sharing—are essential for new technology adoption⁴⁴ . Digital finance is activating the agricultural technological innovation chain through three mechanisms. First, regarding capital supply mechanisms, digital financial tools like internet lending and blockchain financing bypass traditional credit requirements for collateral. They provide initial funding for developing rural-adapted technologies such as drone crop protection and IoT monitoring, effectively addressing insufficient R&D investment in rural innovation processes^45,46. Second, in the information matching mechanism, digital platforms integrate multidimensional data—including soil metrics and market demand signals—to steer R&D toward urgently needed rural sectors like water-saving irrigation and specialty agricultural processing, bridging the gap between technological offerings and actual needs. Third, in risk-sharing mechanisms, digital finance reduces risks during agricultural technology trials through innovative products like weather index insurance and loss compensation for technology application, thereby boosting rural operators’ willingness to adopt new technologies. Through the combined effects of these mechanisms, the efficacy of technological innovation permeates all rural domains: intelligent sorting equipment enhances the added value of agricultural products, driving industrial prosperity; precision irrigation and drone-based crop protection reduce resource consumption and pollution, fostering ecological livability; digital platforms deliver agricultural training, disseminating scientific knowledge to cultivate rural civility; IoT monitoring provides real-time feedback on environmental and production data, enabling effective governance; and increased agricultural productivity directly boosts farmers’ operational income, supporting affluent livelihoods.

From the perspective of capital allocation efficiency pathways, digital finance leverages big data and artificial intelligence to precisely identify pain points in rural capital supply and demand, thereby optimizing capital allocation efficiency across various rural sectors⁴⁷. On one hand, by integrating data such as farmers’ production records, industrial development plans, and market conditions, digital finance enables targeted capital allocation⁴⁸. On the other hand, it reduces capital circulation costs through smart contracts and online trading platforms, accelerating capital concentration in high-value sectors. Simultaneously, the optimization effects of capital allocation extend across multiple dimensions of rural development. Regarding industrial prosperity, capital drives the expansion of industrial chains from production to the entire value chain. In ecological livability, green capital supports rural environmental governance and ecological conservation; In governance effectiveness, digital platforms enable real-time tracking of collective capital flows, enhancing fund management transparency and credibility; In rural cultural advancement, capital supports the development of rural cultural stations and intangible cultural heritage preservation projects, enriching public cultural offerings; In living prosperity, capital precisely empowers farmers to increase their income, drives upgrades in public service facilities such as rural education and healthcare, and improves the quality of people’s lives.

In summary, digital finance drives technological innovation through a three-dimensional mechanism while optimizing capital allocation efficiency through precise resource allocation. Together, these two aspects cover the five dimensions of rural revitalization, forming a multidimensional indirect empowerment system. Based on this, the following research hypothesis is proposed:

H2: Digital finance indirectly empowers rural revitalization by promoting technological innovation.

H3: Digital finance indirectly promotes rural revitalization by enhancing capital allocation efficiency.

Threshold effects of digital finance on rural revitalization

Based on the theory of gradient development⁴⁹, existing disparities in regional economic development levels lead to systemic differentiation in their capacity to absorb and apply new technologies. This pattern is particularly evident in the practice of leveraging digital finance to empower rural revitalization. Specifically, economic development shapes the boundaries of digital finance’s rural empowerment through dual pathways. First, from the supply side of digital infrastructure, economically advanced rural areas possess greater capacity to invest in hardware development—such as network coverage and smart device proliferation—which forms the prerequisite for extending digital financial services. Conversely, economically lagging regions face inherent “hard constraints” on digital finance penetration due to insufficient infrastructure development capabilities. Second, from the demand side perspective of digital literacy, economic development correlates positively with human capital accumulation. As an extension of human capital in the digital era, digital literacy directly influences farmers’ acceptance and proficiency in using digital financial tools. Farmers in economically underdeveloped regions often lack sufficient digital literacy. Even when accessing digital financial services, issues like improper usage and limited functional understanding may diminish their empowering effects.

Beyond this, the level of economic development remains the critical threshold variable determining whether the impact of digital finance transitions from quantitative to qualitative change. Before this threshold is crossed, digital finance faces dual constraints from inadequate infrastructure and low digital literacy, limiting its influence on core areas like rural industrial upgrading and livelihood improvement to localized effects. This explains why some studies find digital finance’s enabling effects on underdeveloped rural areas to be insignificant^50,51. Once the threshold is surpassed, the scale effects of infrastructure and the group effects of digital literacy combine synergistically. On one hand, robust digital infrastructure reduces service costs for financial institutions, incentivizing them to develop tailored financial products for rural areas. On the other hand, the improvement in farmers’ digital literacy accelerates the diffusion of digital finance, forming a positive feedback loop of “use-learn-reuse.” This transforms the role of digital finance in rural revitalization from localized penetration to comprehensive empowerment, resulting in a significant nonlinear leap.

Existing research has largely focused on the direct effects of digital finance on rural revitalization^52,53,54, yet it has rarely incorporated economic development levels into theoretical frameworks, particularly failing to reveal their core role as threshold variables. By deepening the application of gradient development theory, this study not only validates the conditional nature of digital finance’s empowerment for rural revitalization but also bridges the methodological gap in existing literature. It offers a new research perspective for understanding the application boundaries of digital technologies in rural contexts. In summary, this study proposes the following research hypothesis:

H4: The promotional effect of digital finance on rural revitalization exhibits a threshold effect based on economic development levels. Specifically, when economic development crosses a specific threshold, its promotional effect becomes more pronounced.

Model construction and variable selection

Model settings

Benchmark regression model

This paper mainly investigates the impact of digital finance on rural revitalization. Based on the theoretical mechanism analysis, a model with time and regional fixed effects is constructed. To mitigate the impact of heteroscedasticity, logarithmic transformation is applied to all variables. The specific model is as follows:

$${\text{ln}VP}_{it}={\alpha }_{0}+{\alpha }_{1}{\text{ln}DF}_{it}+{\beta }_{1}{\text{X}}_{it}+{\mu }_{i}+{\nu }_{t}+{\varepsilon }_{it}$$

(1)

where, i denotes the city, and t denotes the year. ${\text{ln}VP}_{it}$ is the dependent variable, while ${\text{ln}DF}_{it}$ represents the level of rural revitalization development in prefecture-level cities and serves as the core explanatory variable, reflecting the development level of digital finance. ${\text{X}}_{it}$ includes other control variables that affect rural revitalization; ${\mu }_{i}$ denotes regional fixed effects; ${\nu }_{t}$ represents time fixed effects; ${\varepsilon }_{it}$ is the random disturbance term. ${\alpha }_{1}$ is the key parameter of interest. If ${\alpha }_{1}>0$ is significant, it indicates that an improvement in the level of digital finance can promote the development of rural revitalization.

Mediating effect model

To further explore the mechanisms through which digital finance drives rural revitalization, this section constructs the following model based on the research hypotheses from the theoretical section, using technological innovation and capital allocation efficiency as mediating variables.

$${\text{M}}_{it}={\theta }_{0}+{\theta }_{1}{\text{ln}DF}_{it}+{\beta }_{2}{\text{X}}_{it}+{\mu }_{i}+{\nu }_{t}+{\varepsilon }_{it}$$

(2)

$$\text{ln}{VP}_{it}={\alpha }_{0}+{\alpha }_{1}{\text{ln}DF}_{it}+{\alpha }_{2}{\text{M}}_{it}+{\beta }_{3}{X}_{it}+{\mu }_{i}+{\nu }_{t}+{\varepsilon }_{it}$$

(3)

where, ${\text{M}}_{\text{it}}$ represents the mediating mechanism variable, specifically technological innovation (inn) and capital allocation efficiency (ce), while other variables are identical to those in Eq. (1).

Panel threshold regression model

Considering the potential non-linear impact of digital finance on rural revitalization with increasing marginal effects under different levels of economic development, this paper employs the level of economic development (${pgdp}_{it}$) as a threshold variable to construct a panel threshold regression model:

$${\text{ln}VP}_{it}={\lambda }_{0}+{\lambda }_{1}\text{ln}{DF}_{it}\times d\left(q\le {pgdp}_{it}\right)+{\lambda }_{2}\text{ln}{DF}_{it}\times d\left(q>{pgdp}_{it}\right)+{ \beta }_{4}{X}_{it}+{\mu }_{i}+{\nu }_{t}+{\varepsilon }_{it}$$

(4)

where, $d(*)$ is the indicator function, ${pgdp}_{it}$ is the threshold variable, and ${\lambda }_{1}$ and ${\lambda }_{2}$ represent the elasticity coefficients of the impact of digital finance on rural revitalization at $\text{q}\le {pgdp}_{it}$ and $\text{q}>{pgdp}_{it}$, respectively. If the threshold is properly chosen, the estimates or signs of ${\lambda }_{1}$ and ${\lambda }_{2}$ should differ. Equation (4) only examines the single-threshold effect. Since the analysis of multiple thresholds is logically consistent with that of a single threshold, detailed elaboration is omitted here. This paper will further test and analyze the multiple-threshold effects in the following sections.

Variable selection and explanation

Explained variable

The dependent variable in this study is the level of rural revitalization (VP). Its indicator system construction and measurement methodology are based on the five major requirements outlined in the “Rural Revitalization Strategy Plan (2018–2022)”—namely, “thriving industries, ecological livability, civilized rural culture, effective governance, and prosperous livelihoods.” while drawing upon the research framework proposed by Xu X et al.⁵⁵. Considering the availability of urban-level data, sixteen secondary indicators and thirty specific indicators were established across these five dimensions, forming the evaluation framework for rural revitalization development (see Table 1). During the measurement process, this study first adopted the entropy method employed by Xu X et al. (2022), calculating the rural revitalization development index through a sequence of steps: indicator standardization, indicator weight calculation, entropy and coefficient of variation measurement, and determination of indicator weights. Simultaneously, aligning with the research theme of “Digital Finance Empowering Rural Revitalization” and the prefecture-level analysis context, two optimizations were made to Xu X’s (2022) framework: the research unit was shifted from the provincial level to 281 prefecture-level cities. This avoids the issue in Xu X et al. (2022) where provincial data tends to mask internal urban–rural development imbalances within provinces, enabling more precise capture of digital finance penetration differences across administrative tiers and enhancing the policy practicality of conclusions. Ultimately, this study derived comprehensive rural revitalization indices and five subsystem indices for 281 prefecture-level cities from 2011 to 2022. The comprehensive index ranges from 0 to 1, with higher values indicating greater rural revitalization levels.

Table 1 Evaluation index system for rural revitalization development.

Core explanatory variables

Digital finance (DF) is selected as the core explanatory variable in this paper. Following the approach of most scholars, the city-level index from the Peking University Digital Inclusive Finance Index (2011–2022) is adopted to assess the level of digital finance development in each city.

Control variable

Referring to the studies of multiple scholars^56,57,58, this paper selects control variables such as education level, industrial structure, regional population density, transportation infrastructure level, and fiscal level. The education level (edu) is proxied by the ratio of the number of students in ordinary secondary schools to the total population at the end of the year. The industrial structure (str) is measured by the proportion of the added value of the tertiary industry in the total output value. The transportation infrastructure level (inf) is indicated by the road mileage per square kilometer. The regional population density (pop) is calculated as the ratio of the total population at the end of the year to the administrative area. The fiscal level (gl) is proxied by the local government’s general budget revenue.

Mediating variable

This study selects two mediating variables to systematically reveal the transmission pathways through which digital finance impacts rural revitalization. The first is technological innovation (inn), which adopts the research methodology of Bai M et al. (2023)⁵⁹ by using the number of agricultural technology invention patent applications as a proxy indicator. This metric directly reflects the level of technological R&D and innovation activity in the rural agricultural sector. Second, capital allocation efficiency (ce) is measured using the reciprocal of the capital misallocation index, where higher values indicate greater efficiency. The calculation follows the methodology of Chen YW et al. (2023)⁶⁰ , estimated based on prefecture-level city data on fixed asset stock and labor input.

Threshold variable

This paper selects the level of economic development (pgdp) as the threshold variable. Referring to the study of Wei B H et al. (2023)⁶¹, per capita regional gross domestic product is used as a proxy variable for economic development.

Data source and statistical description

Due to the differences in the establishment years of prefecture-level cities and the limitations of missing statistical data, this paper excludes cities such as Pu’er, Xiangyang, Qinzhou, and Zhongwei. Ultimately, an empirical analysis is conducted on 281 prefecture-level cities and above in China from 2011 to 2022, forming a balanced panel of 3,372 city observations. For sporadic missing values occurring in specific years and cities, the study comprehensively evaluated the compatibility between variable characteristics and interpolation methods, employing linear trend extrapolation and regional mean imputation respectively. Linear trend extrapolation is suitable for variables exhibiting continuous temporal trends. This method estimates values based on the variable’s own time-series characteristics, helping maintain temporal continuity in the data. Regional mean imputation is primarily applied to indicators closely linked to economic development levels. It fills gaps by referencing data from cities with similar economic levels and geographic locations, mitigating potential systematic biases arising from regional heterogeneity. It should be noted that all missing value imputation methods carry inherent assumptions and limitations. To minimize potential bias during interpolation, strict conditions were applied in practice: linear extrapolation was restricted to variables exhibiting clear temporal trends, with extrapolation periods not exceeding two consecutive years. Regional mean imputation employed refined grouping based on per capital GDP and geographic proximity to enhance homogeneity within each group. The data are sourced from the China Statistical Yearbook, China City Statistical Yearbook, China Rural Statistical Yearbook, China Rural Management Statistical Annual Report, China Communications Industry Development Statistical Bulletin, local prefecture-level city statistical yearbooks, and the Digital Inclusive Finance Index Report by the Digital Finance Research Center of Peking University. The descriptive statistics of the specific variables are shown in Table 2.

Table 2 Descriptive statistics of model variables.

China Rural Management and Operations Statistical Yearbook.

Empirical results and analysis

Kernel density estimation

To further explore the dynamic evolution trends of rural revitalization and digital finance development levels across different regions in China, the Kernel density estimation method was employed to plot the kernel density curves for rural revitalization and digital finance, as shown in Figs. 2 and 3.

Fig. 2

Kernel density estimation of rural revitalization, 2011–2022. Note: 1. Curve Shape and Temporal Trends: From 2011 to 2022, the rural revitalization kernel density curve consistently exhibited a multi-peak pattern, with the main peak gradually shifting to the right over the years. This indicates a steady improvement in China’s overall rural revitalization level, with prefecture-level cities at the intermediate development stage (corresponding to the main peak group) serving as the core driving force behind rural revitalization. 2. Multi-Peak and Gradient Effects: A distinct secondary peak emerged after 2016, further differentiating into second and third peaks post-2020. This reflects how some lagging prefecture-level cities gradually formed a second tier through targeted rural revitalization support policies, while a few developed eastern prefecture-level cities—benefiting from early digital financial penetration and robust infrastructure—established a third tier. This ultimately manifests as a “multi-tiered, multi-model” development landscape. 3. Distribution Dispersion: The peak values of the curve show a slight decline over time, indicating that while disparities exist in rural revitalization levels among prefecture-level cities, extreme polarization has not emerged. Overall, the pattern exhibits “steady improvement with graded convergence.”

Fig. 3

Kernel density estimation of digital finance, 2011–2022. Note: 1. Unimodal Distribution and Concentration: Throughout the study period, the digital finance kernel density curve consistently exhibited a unimodal distribution with a high peak, indicating that most prefecture-level cities in China share a similar level of digital finance development without significant polarization. 2. Rightward Shift and Leapfrog Development: 2016 marked a pivotal turning point, with the curve shifting markedly to the right each subsequent year. This corresponds to the implementation of China’s “Broadband China” policy and the expansion of mobile payments into rural areas after 2016, accelerating the improvement of digital financial infrastructure and enabling prefecture-level cities to achieve leapfrog development in digital finance. 3. Right Tail and Regional Leadership: The curve exhibits a pronounced right tail that lengthens over time, indicating that prefecture-level cities in the eastern coastal regions, leveraging their economic foundations and technological advantages, have achieved significantly higher levels of digital finance development compared to other areas. This has fostered a development pattern characterized by “leading cities driving progress while the broader region follows suit.”

The kernel density curve of rural revitalization exhibits a multimodal trend during the study period, as shown in Fig. 2. This intuitively reflects the significant gradient effect in China’s rural revitalization development. Most cities are at the average main peak level, while some lagging cities form the second and third echelons of rural revitalization development, constituting a multi-level and multi-modal development pattern.

The results from Fig. 3 show that the kernel density curve of digital finance exhibits a unimodal distribution during the study period. This indicates that the development of digital finance in China shows a high degree of concentration, meaning that the digital finance levels of most cities are similar, with only a few cities having higher or lower levels of digital finance development. In terms of skewness, the kernel density curve of digital finance has shifted to the right year by year since 2016, indicating that the level of digital finance has increased significantly and achieved leapfrog development after 2016. In terms of distribution extension, the digital finance curve exhibits a rightward tail, suggesting that there are more cities with higher levels of digital finance development.

Analysis of benchmark regression results

In this paper, a double fixed-effects model is used to evaluate the effect of digital finance on rural revitalization. Regression tests are conducted both without control variables and with the sequential inclusion of control variables. Table 3 presents the benchmark regression results of the impact of digital finance on rural revitalization. In the first column, without control variables, the regression results show that digital finance has a significant positive impact on rural revitalization, significant at the 1% level. Columns (2)–(6) show the results with control variables added sequentially, where the coefficients for digital finance remain positive and significant at the 1% level. As control variables are added sequentially, the regression coefficient for digital finance (DF) decreases slightly. Overall, digital finance in China has made rapid progress from 2011 to 2022, fully demonstrating its inclusive nature. It has played a crucial role in promoting rural revitalization and injected strong vitality into the economic and social development of rural areas.

Table 3 Benchmark regression results.

The regression results of the control variables selected in this paper are generally consistent with expectations. The regression coefficient for education level (edu) is positive and significant at the 1% level, indicating that higher regional education levels are more conducive to the development of rural revitalization. Education can enhance farmers’ scientific and cultural literacy, improve their vocational skills, and strengthen their innovation capabilities and market competitiveness, thereby providing a human resource guarantee for promoting rural economic development. Also, the regression coefficient for industrial structure (str) is positive and passes the statistical significance test at the 1% level. This result indicates that as the proportion of the tertiary sector in GDP increases, its role in promoting rural revitalization also strengthens. The development of the service industry means more job opportunities. Against the backdrop of rural labor migration, an increase in urban employment opportunities may lead to higher disposable income for farmers, further improving their living conditions and thereby promoting rural revitalization. Additionally, the regression coefficient for regional population density (pop) is negative and significant at the 5% level, indicating that lower population density in a region is more conducive to rural revitalization. Moreover, lower population density implies reduced per capita resource and environmental pressures, providing a broader development space and better ecological environment for improving rural living conditions. The regression coefficient for transportation infrastructure level (inf) is positive and significant at the 5% level. From this, it can be seen that well-developed transportation facilities can connect rural areas with the outside world, reduce logistics costs, enhance the market competitiveness of agricultural products, promote the upgrading and transformation of rural industries, and drive rural economic development, thereby providing a solid foundation for rural revitalization. The regression coefficient for fiscal level (gl) is positive but does not pass the significance test.

Robust test

Exclude specific samples

The four municipalities, Beijing, Tianjin, Shanghai, and Chongqing, directly under the central government, have higher levels of economic development and more complete network infrastructure, with digital finance development levels far exceeding those of other regions. Referring to the approaches of scholars such as Cao J et al.⁵⁶ and Yao L et al.⁶², the special samples of these four municipalities are excluded to eliminate potential biases in the empirical results due to sample selection, and the promoting effect of digital finance on rural revitalization is re-examined. Column (1) of Table 4 presents the estimation results after excluding specific samples. The results show that after controlling for the interference of special samples, the estimated coefficient of digital finance remains significant at the 1% level. Additionally, the signs of the control variables are consistent with the results of the benchmark regression analysis, further confirming the reliability of the research findings.

Table 4 Results of robustness test.

Change the time span

To exclude the interference of time factors on the empirical results, robustness tests are conducted by changing the time span. Specifically, samples before 2013 are excluded, and the model is re-estimated for different time periods to verify the reliability of the empirical analysis results. As shown in column (2) of Table 4, during the period of 2013–2022, digital finance has a positive impact on rural revitalization and passes the significance test at the 1% level, further enhancing the reliability of the research findings.

Control the development of traditional finance

Digital finance has enriched the scope of financial services and revitalized traditional financial institutions. However, it remains challenging to definitively determine whether the empowerment achieved in rural revitalization stems from the traditional financial system or from digital finance as an emerging financial model. To avoid the estimation effects of digital finance being influenced by traditional financial development, while also considering the sustainability of digital finance effects across different levels of traditional financial development, this paper adds two control variables: traditional financial development (finance) and broadband penetration rate (BPR), and re-estimates the regression results. Traditional financial development (finance) is represented by the ratio of year-end deposits and loans of financial institutions to GDP, following the approach of Zhang L et al.⁵⁷.Broadband penetration rate (BPR) is measured using the number of Internet users per 100 people as a proxy indicator, following the methodology of scholars such as Zhao T et al.⁶³. As shown in column (3) of Table 4, the coefficient indicating the impact of digital finance on rural revitalization is positive and remains statistically significant at the 1% level, further validating the research findings.

Exclusion of samples with high missing data proportion

To examine the potential impact of data missingness on estimation results, a robustness test was further conducted by excluding cities with high missing data proportions. Specifically, if a city’s proportion of missing years in digital finance, rural revitalization indicators, and control variables exceeded 30% of the total observation period, it was classified as a high-missing-proportion sample and excluded. After screening, 12 cities were excluded, leaving a final sample covering 269 cities and 3,228 observations. Results from the re-estimated model (see Column 4 of Table 4) show that the coefficient for digital finance remains significantly positive, consistent in direction and significance level with the benchmark regression results. This indicates that even after excluding samples with poor data integrity, the promotional effect of digital finance on rural revitalization remains robust, and the research conclusions are not systematically influenced by cities with high missing data rates.

Endogeneity test

Instrumental variables method

The benchmark regression results show that digital finance has a significant driving effect on rural revitalization. However, this estimation result may lead to endogeneity issues due to potential bidirectional causality. In light of this issue, this paper refers to the study of Zhao T et al.⁶³, employs the interaction term between the number of fixed telephones per hundred people in prefecture-level cities in 1984 and the number of internet broadband subscribers per ten thousand people nationwide in the previous year as an instrumental variable for digital finance, and uses the two-stage least squares (2SLS) method for quantitative assessment. Table 5 reports the regression results of the instrumental variable 2SLS. The first-stage RKF test statistic is much greater than the empirical value of 10, indicating that the possibility of weak instrumental variables is excluded. Additionally, the coefficient signs and significance of the core explanatory variables in the second-stage regression results are consistent with the benchmark regression results. Moreover, the results remain significant even after including control variables, digital finance significantly promotes rural revitalization. These tests further validate the reliability of the regression results.

Table 5 Results of instrumental variable regression.

Double difference method

This study draws upon the research of Tang K et al.⁶⁴, utilizing the “Broadband China” demonstration cities as a quasi-natural experiment setting. Each city is assigned a value based on the list of “Broadband China” demonstration cities published by China’s Ministry of Industry and Information Technology in 2016. A city is assigned a value of 1 in the year it was designated as a “Broadband China” demonstration city and in subsequent years; otherwise, it is assigned a value of 0. Since the list of demonstration cities was announced in the second half of the year, this paper defines the year following the announcement as the policy implementation year and estimates the policy effect. The analysis is conducted based on the following model:

$${Y}_{it}={\alpha }_{0}+\beta {policy}_{it}+\delta {X}_{it}+{\mu }_{i}+{v}_{t}+{\varepsilon }_{it}$$

(5)

In this equation, ${Y}_{it}$ is the dependent variable, representing the rural revitalization level of city $i$ in year t. ${v}_{t}$ represents the fixed time effect, ${\mu }_{i}$ denotes the individual fixed effect for each city, and ${\varepsilon }_{it}$ is the random error term. ${X}_{it}$ is a series of variables that may impact rural revitalization. ${policy}_{it}$ is the core explanatory variable, representing a dummy variable for “Broadband China” demonstration cities. Its coefficient $\beta$ measures the impact of “Broadband China” demonstration city construction on rural revitalization. If $\beta$ is positive and significant, it indicates that the construction of “Broadband China” demonstration cities can promote rural revitalization. Conducting policy impact assessments using the difference-in-differences method requires the precondition that the control group and experimental group share a common trend. Therefore, this paper employs a dynamic difference-in-differences approach to test for parallel trends, with results shown in Fig. 4. Additionally, although the regression model controls for a series of urban characteristics affecting rural revitalization, unobservable factors that vary over time and location may still exist. These could influence estimation results and introduce estimation errors. Therefore, this study employs an indirect placebo test by randomly selecting pilot cities from the “Broadband China” initiative. Based on the regression model, 500 simulated regressions were repeated, with results shown in Fig. 5.

Fig. 4

Parallel trend test results.

Fig. 5

Placebo test results.

Figure 4 results indicate that prior to the implementation of the “Broadband China” pilot policy, no significant systemic differences existed between pilot and non-pilot cities in terms of rural revitalization. Following policy implementation, however, significant differences emerged between the two groups, satisfying the parallel trend assumption and confirming the validity and appropriateness of the difference-in-differences approach. Figure 5 results show that the P-value distribution from 500 simulated regressions and the kernel density distribution of regression coefficients both exhibit near-normal distributions, indicating that the regression results are free from endogeneity issues.

Further analysis

Analysis of impact mechanism

By constructing a basic regression model and conducting robustness tests through methods such as excluding specific samples, altering the time span, controlling for traditional financial development, and accounting for other omitted variables, it can be confirmed that digital finance promotes rural revitalization. Subsequently, to examine whether technological innovation and capital allocation efficiency serve as mediating factors in the process of digital finance influencing rural revitalization, these variables were treated as mediators and validated using a mediation effect model. Detailed results are presented in Table 6.

Table 6 Results of mediating effect test.

Table 6 reports the mechanism through which digital finance influences rural revitalization via two mediating pathways: technological innovation and capital allocation efficiency.

First, regarding the technological innovation pathway, Column (1) results show that the total effect coefficient of digital finance on rural revitalization is 0.1412, significant at the 1% level, indicating that digital finance development has a significant direct promotional effect on rural revitalization. Column (2), with technological innovation as the dependent variable, shows a coefficient of 0.3393 for digital finance, also significant at the 1% level, indicating that digital finance can effectively enhance regional technological innovation levels. Crucial evidence emerges in Column (3): when both digital finance and technological innovation are introduced as explanatory variables, the coefficient for technological innovation is 0.0177 and passes the 1% significance test.Meanwhile, the coefficient for digital finance decreased from 0.1412 in Column (1) to 0.1354. Although still statistically significant, this reduction indicates that technological innovation has assumed a partial intermediary role in the process of digital finance influencing rural revitalization. This finding confirms that digital finance, through technologies such as big data and cloud computing, fosters a safer and more efficient financing environment for technological innovation activities. Consequently, by enhancing agricultural production efficiency and optimizing industrial structures, it injects new momentum into rural revitalization.

Second, the analytical logic follows a similar path regarding capital allocation efficiency. Column (4) examines the impact of digital finance on capital allocation efficiency, yielding a coefficient of 0.1521 that is statistically significant at the 1% level. This confirms that the development of digital finance significantly enhances capital allocation efficiency. Column (5) incorporates both digital finance and capital allocation efficiency into the regression model. The results show that the coefficient for capital allocation efficiency is 0.0289 and highly significant. Correspondingly, the coefficient for digital finance decreases from 0.1412 in Column (1) to 0.1368 in Column (5). This change further confirms the partial mediating effect of capital allocation efficiency. It indicates that digital finance, leveraging its informational advantages and risk control capabilities, guides capital more effectively toward critical sectors and underdeveloped areas in rural regions. This optimizes the spatial allocation of financial resources, thereby indirectly accelerating the rural revitalization process. A comparative analysis of the mediating effect values across both pathways reveals that technological innovation exerts a relatively stronger mediating role.

Table 6 measurement results only indicate that technological innovation and capital allocation efficiency mediate the relationship between digital finance and rural revitalization, while the magnitude of this mediation effect requires further quantification. Therefore, this paper incorporates Bootstrap-based confidence intervals and effect size metrics. Using STATA 17.0 software, the Bootstrap method was employed to perform 1,000 repeated samples, calculating bias-corrected confidence intervals for the mediation effect. The results are presented in Table 7.

Table 7 Bootstrap mediating effect test results.

Threshold Effect Analysis

The economic development level is selected as a threshold variable and tested based on a panel threshold model to examine whether there are non-linear characteristics of the impact of digital finance on rural revitalization. In this paper, STATA18.0 software and Bootstrap method of repeated bootstrap sampling 300 times were used to test the threshold effect of using the level of economic development as the threshold variable. Table 8 presents the results of the threshold effect test based on Bootstrap method.

Table 8 Threshold effect Self-sampling test.

The results in Table 8 show that the tests for single and double thresholds with the economic development level as the threshold variable are significant at the 1% level, while the triple threshold does not pass the significance test. This indicates that the level of economic development has a double-threshold effect in the impact of digital finance on rural revitalization. After the self-sampling test for the threshold effect, the thresholds of the panel threshold model are estimated and tested, and the results are shown in Table 9.

Table 9 Threshold estimation results.

According to the estimation method proposed by Hansen, the threshold is the value of γ corresponding to the likelihood ratio statistic LR when it tends to zero. Therefore, this section plots the LR plots of the corresponding threshold estimates at 95% confidence intervals (Fig. 6). This visually demonstrates the process of constructing the estimates and confidence intervals for the two thresholds of economic development. In the LR plots, the dashed lines below refer to the threshold intervals that correspond to the critical value of 7.35 at the LR value of less than the 5% level of significance.

Fig. 6

Dual threshold estimation results and confidence intervals. Note: 1. The horizontal axis represents the estimated range of threshold values, while the vertical axis shows the likelihood ratio statistic (LR). The dashed line indicates the critical value at the 5% significance level (LR = 7.35). According to Hansen’s threshold estimation theory, when the LR value is below the critical threshold, the corresponding threshold value lies within the 95% confidence interval, indicating a valid threshold estimate; 2. First Threshold: The LR value approaches 0 near the estimated value of 3.2012, with a narrow 95% confidence interval [3.1452, 3.2130], indicating high estimation precision for the first threshold. When the economic development level (pgdp) of a prefecture-level city crosses 3.2012, the structural change in digital finance’s promotion of rural revitalization first emerges; 3. Second Threshold: The LR value approaches 0 near the estimated value of 7.1087, with a similarly narrow 95% confidence interval [7.0243, 7.1511]. is similarly narrow, validating the robustness of the second threshold. This indicates that beyond an economic development level of 7.1087, the promotional effect of digital finance further intensifies, exhibiting a “marginal increase” characteristic.

According to Table 9 and Fig. 6, the estimated values of the double threshold of economic development are distributed as 3.2012 and 7.1087, and the corresponding confidence intervals are [3.1452,3.2130] and [7.0243,7.1511], respectively. Therefore, the economic development level of 281 cities in China during the study period can be classified into three levels: regions with low level of economic development (pgdp < 3.2012), regions with medium level of economic development (3.2012 ≤ pgdp ≤ 7.1087) and regions with high level of economic development (pgdp > 7.1087).

Evidently, there are differences in the impact of Chinese digital finance on rural revitalization under different levels of economic development during the study period, as shown by the results of the double-threshold parameter estimation based on the level of economic development (see Table 10). According to the regression results, there is a significant positive impact of digital finance on rural revitalization in the low level of economic development interval (pgdp < 3.2012), with a coefficient of 0.1041, and it passes the test of significance at the 1% level; In the interval of medium level of economic development (3.2012 ≤ pgdp ≤ 7.1087), the regression coefficient of digital finance is 0.1132, which is significant at the 1% level. It indicates that digital finance has further increased its contribution to the development of rural revitalization in this interval; In the interval of high level of economic development (pgdp > 7.1087), digital finance significantly pushes the process of rural revitalization, and its impact coefficient is as high as 0.1244, and it passes the test at 1% level of significance. This indicates that when the level of economic development crosses the threshold, the driving effect of digital finance on the development of rural revitalization appears to have a non-linear characteristic of marginal increment, thus verifying Hypothesis 3.

Table 10 Estimation results of dual threshold parameters for financial development.

Heterogeneity analysis

Differences in the level of digital finance development and the process of rural revitalization in different geographical regions lead to inconsistencies in the degree of impact of digital finance on the level of rural revitalization development in different regions. Therefore, a dual fixed-effects model based on individual and time is used to test whether digital finance only contributes to the development of rural revitalization in some regions. From the perspective of the heterogeneity test, the impact of digital finance on rural revitalization in the three major regions of eastern, central and western China is investigated based on China’s traditional geographic divisions. The results of correlation regression analysis are shown in Table 11. The data show that in the eastern region, the promotion coefficient of digital finance on rural revitalization is 0.1914 and passes the significance test at the 1% level; the effect coefficient in the central region is -0.0137, but not significant; while in the western region, the effect coefficient is 0.1352, which passes the test of significance level of 1%. Overall, the promoting effect of digital finance shows significant regional differences during rural revitalization: it has a significantly positive impact in the eastern region, while in the central region, the development of digital finance has a negative but insignificant effect on rural revitalization. In the western region, although digital finance can promote rural revitalization, there is a gap compared with the eastern region. The possible reason is that the eastern region has more complete digital infrastructure, such as high-speed networks and mobile payment systems, which provide a foundation for the widespread application and in-depth penetration of digital financial services into rural areas. Moreover, the eastern region has a stronger economic foundation and robust financial demand. Farmers in this region have a relatively higher acceptance and usage rate of digital finance, which can effectively provide sufficient financial support for rural revitalization. Compared with the developed eastern coastal regions, the central region still has a gap in digital infrastructure construction. Although its hardware facilities may be more complete than those in the western region, the central region has a higher population density and a more pronounced urban–rural dual structure. A large number of young and middle-aged laborers work outside, leaving a higher proportion of elderly people and children in the local area. These groups have relatively weaker acceptance of new technologies, which reduces the penetration and usage frequency of digital financial products and services, and thus the promoting effect on rural revitalization is not significant. In contrast, in some western regions, the insufficient coverage of traditional financial services is more prominent due to geographical constraints. Under such circumstances, digital finance has become an effective means to solve the “last mile” problem and has thus promoted its rapid popularization and in-depth application. Additionally, supported by the national Western Development Strategy, green industries such as eco-tourism and the processing of specialty agricultural products have received focused support in the western region, offering greater potential for integration and innovation with digital finance. Therefore, western regions are likely to show higher adaptability and flexibility in promoting the upgrading of the agricultural industry and fostering new forms of agricultural business, enabling digital finance to better serve rural revitalization.

Table 11 Heterogeneity analysis results in different regions.

Conclusion and countermeasures

The close integration of digital finance and rural revitalization is an important direction of current financial reform and rural development strategy in China. In this paper, 281 cities panel data in China from 2011 to 2022 are collected to explore the impact and mechanism of digital finance on rural revitalization through using individual and time dual-way fixed effect model, mediated effect model and threshold effect model. The following conclusions are drawn:

• (1) The basic regression results indicate that digital finance significantly promotes rural revitalization. This conclusion remains valid after a series of robustness tests, including excluding specific samples, changing the time span, controlling the traditional financial development, considering other omitted variables, and addressing endogeneity issues.

• (2) The results of the mechanism testing indicate that the enabling effect of digital finance on rural revitalization is not direct, but rather achieved indirectly through two intermediary pathways: technological innovation and capital allocation efficiency. Through the technological innovation pathway, digital finance overcomes constraints such as funding shortages, information asymmetry, and high risks in rural R&D, facilitating the adoption of suitable technologies and driving rural revitalization through innovation. Via the capital allocation efficiency pathway, digital finance optimizes resource allocation by precisely matching rural capital supply and demand, reducing factor transaction costs, and enhancing capital efficiency in agricultural production and rural services, thereby strengthening the foundational elements for rural revitalization.

• (3) The threshold effect test results show that there are double thresholds in the process of digital finance affecting rural revitalization. The first and second threshold values are 3.2012 and 7.1087, respectively. In the process of crossing the two thresholds, that is, with the improvement of the level of economic development, the promoting effect of digital finance on rural revitalization shows a non-linear characteristic of increasing marginal returns.

• (4) The heterogeneous results indicate that the impact of digital finance on rural revitalization varies across different geographical regions. Specifically, digital finance significantly promotes rural revitalization in the eastern region, while it has an insignificant impact on rural revitalization and may even have a negative effect in the central region. In the western region, although digital finance can promote rural revitalization, its effect is not as strong as that in the eastern region. These differences are mainly influenced by factors such as regional economic development levels, digital infrastructure, talent, and technology.

Based on the above conclusions, this paper proposes the following countermeasures:

First, efforts should be made to advance the rapid development of digital finance and form a deep synergy with the rural revitalization strategy. Specifically, it is necessary to focus on expanding the coverage of digital financial services and accurately directing financial resources towards rural industries to empower rural industrial revitalization with modern financial services. Firstly, the rapid development opportunity of digital information technology should be fully seized to improve the digital financial infrastructure in rural areas, including network and power facilities, to ensure the stability and security of digital financial services and provide a solid hardware foundation for the close integration of digital finance and rural revitalization. Secondly, financial institutions should be encouraged to expand their business in rural areas, enrich digital financial products, and meet the diversified financial needs of rural areas. For example, financial products tailored to the characteristics of rural areas, such as agricultural insurance and rural microloans, should be developed to precisely meet the needs of agricultural production and rural life and provide more comprehensive and personalized financial services. Finally, the digital financial literacy of farmers should be improved to enhance the depth of digital finance usage. For instance, farmers’ understanding of digital finance and their ability to use digital financial tools such as mobile payments, online banking, and digital credit can be enhanced by conducting digital financial knowledge education and practical training. Besides, financial risk education should also be carried out to make farmers aware of potential risks in digital finance, such as online fraud and information security, and to enhance their risk prevention awareness.

Second, the level of technological innovation should be enhanced to serve rural revitalization in a more efficient and precise manner. Firstly, technology incubation platforms should be established to attract and nurture agricultural technology companies, promoting the transformation and application of agricultural scientific and technological achievements. Second, governments, enterprises, and society should jointly establish agricultural technology innovation funds through subsidies, investments, loans, and other means to provide financial support for agricultural technology research and innovation projects. Third, digital technology education and training should be conducted to improve the digital technology literacy of rural residents, and to cultivate a group of new farmers who understand agriculture, technology, and the market, thereby providing a talent guarantee for rural revitalization.

Third, optimize the efficiency of rural capital allocation to lay a solid foundation for empowering rural revitalization through digital finance. First, establish a platform for precise urban–rural capital matching, integrating the needs of rural specialty industries, farmers’ business plans, and information on idle urban capital. Through data-driven matching, guide capital toward high-potential sectors such as organic agriculture and rural cultural tourism, preventing blind investment in inefficient industries. Second, improve the rural capital circulation service system. Leverage digital tools to streamline processes for land management rights mortgages and agricultural supply chain financing, reducing intermediary costs like appraisal and guarantees to enhance capital flow efficiency in rural production and operations. Third, establish a dynamic monitoring mechanism for capital allocation. Track the usage and effectiveness of rural collective capital and credit funds in real time, issuing timely warnings and adjustments for inefficiently allocated projects.

Fourth, region-specific and differentiated economic development strategies should be implemented. The level of economic development is a crucial factor in enabling digital finance to boost rural revitalization. Based on a thorough understanding of their natural conditions, resource endowments, industrial structures, population structures, and socio-cultural characteristics, each region needs to formulate personalized economic development strategies. While promoting their own development, regions also need to enhance cooperation among each other. This can be achieved through various cooperative models, such as signing inter-regional strategic cooperation agreements and jointly investing in infrastructure projects like transportation and communication, to improve connectivity. At the same time, non-governmental organizations, industry associations, and enterprises should be encouraged to participate in regional cooperation, forming a diversified cooperative ecosystem and establishing mutually beneficial regional cooperative relationships.

Discussion

International relevance

Empirical analysis of panel data from China’s 281 prefecture-level cities demonstrates that digital finance exerts a significant direct positive impact on rural revitalization. Technological innovation and capital allocation efficiency serve as core mediating factors. Economic development faces dual thresholds, with pronounced regional heterogeneity. This conclusion provides a “gradient adaptation”framework for international applications.

Low-income countries such as Ethiopia and Nepal can draw on China’s western development experience by prioritizing low-cost digital infrastructure like 4G base stations and affordable networks, while promoting lightweight tools like digital payments to bridge traditional financial gaps. Scenario-based training should be implemented to enhance farmers’ digital literacy and avoid the “usage gap.” Middle-income countries like Thailand and Mexico can adopt China’s “digital finance + technological innovation” model by launching initiatives such as “smart agricultural machinery loans” and “agricultural technology incubation funds,” linking digital credit with agrotechnology adoption. They should promote urban–rural digital resource coordination to foster complementarity between traditional and digital finance, avoiding constraints similar to China’s central region urban–rural dual structure. High-income countries like Australia and Canada can deepen China’s “industrial integration” approach by building blockchain-based “agricultural product traceability-credit” systems, launching “digital credit for eco-farms” that incorporates ecological value into risk management; and integrating digital finance with rural governance to achieve intelligent public services.

Challenges in policy implementation

At the digital finance advancement level, weak network and power infrastructure in western rural areas, coupled with insufficient digital literacy among left-behind populations, result in low penetration rates for tools like mobile payments and digital credit. At the technological innovation enhancement level, agricultural technology incubation platforms are scarce in central and western regions, while capital and technical talent concentrate in the east, leaving technology commercialization lacking practical implementation scenarios. Regarding differentiated strategies, regional cooperation faces barriers in coordinating interests. For instance, resource complementarity between eastern and central/western regions is weak, and disparities in local fiscal capacity make it difficult for low-income areas to bear digital infrastructure investments. Acknowledging these obstacles enhances the feasibility of recommendations and prevents policy measures from becoming overly abstract.

Research limitations and future directions

Although this study systematically reveals the impact and mechanisms of digital finance on rural revitalization based on panel data from 281 prefecture-level cities, there remains room for expansion. First, in terms of data dimensions, the existing analysis primarily relies on macro-level data at the prefecture-level city level and has not incorporated micro-level data such as household micro-level decision-making or the behavior of agricultural business entities. This makes it difficult to accurately capture the differences in micro-level transmission pathways through which digital finance influences rural revitalization. Second, regarding variable measurement, the proxied indicator for technological innovation—the number of agricultural technology invention patent applications—emphasizes the R&D phase. It does not fully capture subsequent stages such as technology transfer and practical implementation, potentially limiting the comprehensiveness of the mechanism analysis.

Future research can deepen in three areas: First, integrate micro-level survey data to explore the differentiated impacts of digital finance on various types of farmers and agricultural enterprises. Second, expand the mechanism analysis framework by introducing moderating variables such as digital infrastructure and human capital to further unravel the enabling logic under multi-factor interactions. Third, examine the dynamic effects of digital currencies like the digital yuan and green digital finance on rural revitalization, providing more timely empirical support for policy optimization.